We collected nine saliva samples throughout the day: Five samples were taken in the morning in order to assess the Cortisol Awakening Response (CAR) and four samples were taken in the afternoon in order to assess the Diurnal Baseline Cortisol (DBC) levels. After 8 months, a follow-up measurement was conducted, repeating all the measures as at t1. We previously identified a flattened CAR in patients with functional neurological disorders (FND) in comparison to healthy controls (HC).

The exact procedure of the saliva sampling can be found here: Weber - 2022 - Identification of biopsychologial trait markers in functional neurological disorders

1 Introduction

In the previous work, we could show that FND patients have a flatter cortisol awakening response (CAR) compared to healthy controls (HC), Figure 1.

Figure 1

Also, we could show that the reduced CAR was associated not only with severity but also duration of past emotional neglect (Figure 2).

This suggests some sort of long-term maladaptive habituation of the HPA-axis in FND as a response to prolonged emotional stress/neglect!

Figure 2

Additionally, we could show that FND patients had a significantly reduced hippocampal (and amygdalar) volume (Figure 3).

Figure 3

In healthy controls, we found that higher cortisol levels were associated with smaller brain volumes. This would support a neurotoxicity hypothesis. Due to the absence of such an association in FND, we suggest that reduced hippocampal volume potentially represents a biological vulnerability factor for FND in the form of a stress-diathesis model, which refers to having a predisposition to a condition which might be activated through stress.

1.1 Study design and aim of the current project

Based on previous results, we aim at investigating the natural course of FND. As such, 53 FND patients participated in a follow-up measurement. At both timepoints (T1 and follow-up(FUP)), the following measures were done:

Neuroimaging (structural, resting-state fMRI)

Cortisol samples

Questionnaires (general health, mood)

Neurological examination

Only at t1, we assessed the following data:

Genetics/epigenetics of certain genes associated with the stress pathway

Data on menstrual cycle

healthy control cohort

2 Demographic and Clinical Data

In total, 53 FND patients participated in the follow-up measurement. We can first look at demographic and clinical characteristics.

	Total	t1	fup	P-value
Demographic and clincal characteristics
Age, mean (SD)	38 (±15)	38 (±15)	38 (±15)	1.00
Sex [females]	76 (72%)	38 (72%)	38 (72%)	1.0
BDI - Depression, mean (SD)	13 (±10)	13 (±10)	12 (±11)	0.66
STAI1 - State Anxiety, mean (SD)	36 (±11)	35 (±11)	36 (±11)	0.60
S-FMDRS - Symptom Severity, mean (SD)	8 (±10)	8 (±9)	9 (±11)	0.65
Subjective Symptom Severity, mean (SD)	41 (±29)	43 (±29)	39 (±29)	0.45

Patients did not differ in their clinical characteristics from T1 to Follow-up. We can also look at the plots and the individual development.

3 Cortisol in FND

Previously, we analysed two metrics to assess cortisol levels: the cortisol awakening response (CAR) and the diurnal baseline cortisol (DBC). The CAR describes the rapid increase in cortisol secretion across the first 30–45 min upon awakening. To assess cortisol differences between T1 and Follow-up in the CAR, a repeated-measures ANOVA was used on the fitted data of the five morning samples (wake-up until 60 min post-awakening) using a linear mixed model with fixed effects of factor Session (T1 vs. FUP) and time point, and using age, sex, symptom severity, smoking, hormonal contraception, psychotropic medication, and menopause as covariates of no interest.
The DBC represents the dynamic changes of cortisol throughout the afternoon (from 2 p.m. to 5 p.m.). As previously, no changes in DBC were found, I did not analyse DBC in the longitudinal data.
For the analyses of the CAR, I excluded data from one FND patients it did not properly adhere to the saliva sampling protocol (delayed sampling with a strict sampling accuracy margin of Δt > 5 min for post-awakening samples).

First we visualize our data.

3.1 Cortisol: Visualization

We can already see that the values are relatively stable (at FUP in comparison to T1). Next we run the statistics.

3.2 Cortisol: Statistics

In our model, we can see that there is no significant effect of Session - no differences in cortisol between T1 and FUP. However, there seems to be a significant effect of symptom severity (S-FMDRS) on the cortisol.

There is a significant association between changes in symptom severity and changes in cortisol.

Visually and statistically, we cannot detect any differences between the measurement timepoints, nor any interactione effects between Session (t1 vs. follow-up) and symptom severity (sfmdrs).

This means:

No significant effect of Session: Patients do not differ between T1 and FUP regarding their CAR levels

No significant interaction effect: Patients who improved/worsened in symptom severity did not change in their cortisol levels

Significant effect of timepoint: This is expected, as the CAR increases and decreases over time

So, how exactly are changes in symptom severity associated with the (non-visible) changes in the CAR without a (session) interaction effect?

##                   Df Sum Sq Mean Sq F value   Pr(>F)    
## timepoint          4  241.6   60.40  10.718 2.41e-08 ***
## Session            1    0.0    0.00   0.000 0.983292    
## sfmdrs             1   79.2   79.24  14.060 0.000198 ***
## DurationSymptoms   1   28.5   28.47   5.051 0.025036 *  
## gender             1   31.3   31.28   5.551 0.018855 *  
## smoke              1    0.0    0.00   0.001 0.978562    
## psychMed           1   22.6   22.60   4.011 0.045747 *  
## menopause          1    1.4    1.44   0.256 0.613001    
## contraception      1   11.3   11.25   1.997 0.158213    
## age                1   14.1   14.14   2.510 0.113779    
## Session:sfmdrs     1    9.1    9.13   1.619 0.203760    
## Residuals        505 2845.9    5.64                     
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

3.2.1 Delta Scores

In order to reduce the dimensionality of our data, we can calculate delta scores, i.e., subtracting t1 from fup data. This will reduce our model by one full factor (Session).
We again plot our data, and run a repeated-measures ANOVA was used on the fitted data of the five morning samples (wake-up until 60 min post-awakening) using a linear mixed model with fixed effects of time point, and using age, sex, symptom severity, smoking, hormonal contraception, psychotropic medication, and menopause as covariates of no interest.
(Caveat: Factor Session is removed, as we work with Delta Scores)

We again detect a significant effect of symptom severity. This confirms: A change in symptom severity is associated with a change in CAR.

##                        Df Sum Sq Mean Sq F value  Pr(>F)    
## timepoint               4   18.4    4.60   0.678 0.60801    
## deltaSFMDRS             1   50.7   50.70   7.464 0.00676 ** 
## deltaBDI                1  101.7  101.71  14.975 0.00014 ***
## deltaSTAI               1    0.0    0.02   0.002 0.96045    
## gender                  1    0.0    0.02   0.003 0.95667    
## smoke                   1    0.6    0.64   0.095 0.75863    
## contraception           2   69.6   34.80   5.124 0.00662 ** 
## age                     1   12.0   12.04   1.773 0.18425    
## psychMed                1   36.2   36.15   5.323 0.02190 *  
## timepoint:deltaSFMDRS   4   40.3   10.07   1.482 0.20815    
## Residuals             242 1643.7    6.79                    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

3.3 Stratification based on clinical course at FUP (S-FMDRS)

In order to disentangle the effect of symptom severity on the CAR, we stratify our patients into subgroups based on the S-FMDRS score. As such, we created a delta-score by subtracting the t1 score from the fup score.
This leaves us with:

Positive Delta SFMDRS Score: Patient got worse

No change in Delta SFMDRS Score: Patient did not improve, nor worsen

Negative Delta SFMDRS Score: Patient improved.

We now look at patients who improved in comparison to those who remained or worsened in their symptom severity.

It seems that patients that improved had generally a lower S-FMDRS score as patients that remained or worsened. Using the ANOVA (Chapter 2), we did not detect any differences in symptom severity between time points. Let’s do baseline and FUP comparisons, as well as another ANOVA with another factor (clinical course).

Statistical test
We detect a significant effect of group (F(1,950) = 7.51, P = 0.006), as well as a significant interaction effect between group and session (F(1,950) = 25.09, P > 0.0001).
This means that:

Patients who were initially (t1) high in symptom severity improved for fup

Patients who were initially (t1) low(er) got worse at fup.

##                Df Sum Sq Mean Sq F value   Pr(>F)    
## Session         1    188   187.6   1.981  0.15963    
## group           1    711   711.1   7.510  0.00625 ** 
## Session:group   1   2376  2375.6  25.089 6.53e-07 ***
## Residuals     950  89951    94.7                     
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

## 
##   Simultaneous Tests for General Linear Hypotheses
## 
## Multiple Comparisons of Means: User-defined Contrasts
## 
## 
## Fit: glm(formula = sfmdrs ~ groups + bdi + stai1, family = gaussian)
## 
## Linear Hypotheses:
##                                                                    Estimate
## (worse.fup) - (worse.t1) == 0                                        3.6436
## (improved.fup) - (improved.t1) == 0                                 -3.6029
## (worse.fup) - (improved.fup) == 0                                    6.4895
## (worse.t1) - (improved.t1) == 0                                     -0.7570
## ((worse.fup) - (worse.t1)) - ((improved.fup) - (improved.t1)) == 0   7.2464
##                                                                    Std. Error
## (worse.fup) - (worse.t1) == 0                                          0.7500
## (improved.fup) - (improved.t1) == 0                                    1.0832
## (worse.fup) - (improved.fup) == 0                                      0.9478
## (worse.t1) - (improved.t1) == 0                                        0.9346
## ((worse.fup) - (worse.t1)) - ((improved.fup) - (improved.t1)) == 0     1.3182
##                                                                    z value
## (worse.fup) - (worse.t1) == 0                                        4.858
## (improved.fup) - (improved.t1) == 0                                 -3.326
## (worse.fup) - (improved.fup) == 0                                    6.847
## (worse.t1) - (improved.t1) == 0                                     -0.810
## ((worse.fup) - (worse.t1)) - ((improved.fup) - (improved.t1)) == 0   5.497
##                                                                    Pr(>|z|)    
## (worse.fup) - (worse.t1) == 0                                      1.18e-06 ***
## (improved.fup) - (improved.t1) == 0                                0.000881 ***
## (worse.fup) - (improved.fup) == 0                                  7.55e-12 ***
## (worse.t1) - (improved.t1) == 0                                    0.417955    
## ((worse.fup) - (worse.t1)) - ((improved.fup) - (improved.t1)) == 0 3.86e-08 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- none method)

	Total	improved	worsened/remained	P-value
Patient characteristics at T1, based on later clinical improvement
Age, mean (SD)	38 (±15)	44 (±18)	35 (±12)	0.057
Sex [females]	38 (72%)	12 (71%)	26 (72%)	1.0
BDI - Depression, mean (SD)	13 (±10)	15 (±10)	12 (±10)	0.36
STAI1 - State Anxiety, mean (SD)	35 (±11)	35 (±10)	35 (±11)	0.94
S-FMDRS - Symptom Severity, mean (SD)	8 (±9)	9 (±4)	8 (±11)	0.57
Subjective Symptom Severity, mean (SD)	43 (±29)	50 (±27)	41 (±30)	0.30

	Total	improved	worsened/remained	P-value
Patient characteristics based on clinical improvement at FUP
Age, mean (SD)	38 (±15)	44 (±18)	35 (±12)	0.057
Sex [females]	38 (72%)	12 (71%)	26 (72%)	1.0
BDI - Depression, mean (SD)	12 (±11)	14 (±9)	12 (±11)	0.49
STAI1 - State Anxiety, mean (SD)	36 (±11)	34 (±10)	38 (±11)	0.26
S-FMDRS - Symptom Severity, mean (SD)	9 (±11)	5 (±4)	11 (±13)	0.099
Subjective Symptom Severity, mean (SD)	39 (±29)	39 (±27)	39 (±31)	0.97

3.3.1 Stratification based on other clinical scores

For completedness, we can also look at differences in BDI and STAI.

Statistical test

## [1] "Results BDI:"

##                Df Sum Sq Mean Sq F value   Pr(>F)    
## Session         1    180   179.7   1.755 0.185615    
## group           1   1234  1233.7  12.048 0.000542 ***
## Session:group   1     11    10.7   0.105 0.746315    
## Residuals     950  97273   102.4                     
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

## 
##   Simultaneous Tests for General Linear Hypotheses
## 
## Multiple Comparisons of Means: User-defined Contrasts
## 
## 
## Fit: glm(formula = bdi ~ groups + stai1 + sfmdrs, family = gaussian)
## 
## Linear Hypotheses:
##                                                                    Estimate
## (worse.fup) - (worse.t1) == 0                                       -2.5409
## (improved.fup) - (improved.t1) == 0                                  0.2413
## (worse.fup) - (improved.fup) == 0                                   -5.2781
## (worse.t1) - (improved.t1) == 0                                     -2.4959
## ((worse.fup) - (worse.t1)) - ((improved.fup) - (improved.t1)) == 0  -2.7822
##                                                                    Std. Error
## (worse.fup) - (worse.t1) == 0                                          0.6406
## (improved.fup) - (improved.t1) == 0                                    0.9268
## (worse.fup) - (improved.fup) == 0                                      0.8080
## (worse.t1) - (improved.t1) == 0                                        0.7912
## ((worse.fup) - (worse.t1)) - ((improved.fup) - (improved.t1)) == 0     1.1356
##                                                                    z value
## (worse.fup) - (worse.t1) == 0                                       -3.966
## (improved.fup) - (improved.t1) == 0                                  0.260
## (worse.fup) - (improved.fup) == 0                                   -6.532
## (worse.t1) - (improved.t1) == 0                                     -3.155
## ((worse.fup) - (worse.t1)) - ((improved.fup) - (improved.t1)) == 0  -2.450
##                                                                    Pr(>|z|)    
## (worse.fup) - (worse.t1) == 0                                      7.30e-05 ***
## (improved.fup) - (improved.t1) == 0                                 0.79458    
## (worse.fup) - (improved.fup) == 0                                  6.48e-11 ***
## (worse.t1) - (improved.t1) == 0                                     0.00161 ** 
## ((worse.fup) - (worse.t1)) - ((improved.fup) - (improved.t1)) == 0  0.01428 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- none method)

## [1] "Results STAI-1:"

##                Df Sum Sq Mean Sq F value  Pr(>F)   
## Session         1    286   285.6   2.518 0.11287   
## group           1    800   799.8   7.052 0.00805 **
## Session:group   1    611   611.2   5.389 0.02047 * 
## Residuals     950 107748   113.4                   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

## 
##   Simultaneous Tests for General Linear Hypotheses
## 
## Multiple Comparisons of Means: User-defined Contrasts
## 
## 
## Fit: glm(formula = stai1 ~ groups + bdi + sfmdrs, family = gaussian)
## 
## Linear Hypotheses:
##                                                                    Estimate
## (worse.fup) - (worse.t1) == 0                                        3.0980
## (improved.fup) - (improved.t1) == 0                                 -1.0375
## (worse.fup) - (improved.fup) == 0                                    5.8364
## (worse.t1) - (improved.t1) == 0                                      1.7008
## ((worse.fup) - (worse.t1)) - ((improved.fup) - (improved.t1)) == 0   4.1356
##                                                                    Std. Error
## (worse.fup) - (worse.t1) == 0                                          0.6825
## (improved.fup) - (improved.t1) == 0                                    0.9894
## (worse.fup) - (improved.fup) == 0                                      0.8617
## (worse.t1) - (improved.t1) == 0                                        0.8477
## ((worse.fup) - (worse.t1)) - ((improved.fup) - (improved.t1)) == 0     1.2093
##                                                                    z value
## (worse.fup) - (worse.t1) == 0                                        4.539
## (improved.fup) - (improved.t1) == 0                                 -1.049
## (worse.fup) - (improved.fup) == 0                                    6.773
## (worse.t1) - (improved.t1) == 0                                      2.006
## ((worse.fup) - (worse.t1)) - ((improved.fup) - (improved.t1)) == 0   3.420
##                                                                    Pr(>|z|)    
## (worse.fup) - (worse.t1) == 0                                      5.65e-06 ***
## (improved.fup) - (improved.t1) == 0                                0.294349    
## (worse.fup) - (improved.fup) == 0                                  1.26e-11 ***
## (worse.t1) - (improved.t1) == 0                                    0.044811 *  
## ((worse.fup) - (worse.t1)) - ((improved.fup) - (improved.t1)) == 0 0.000627 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- none method)

We can indeed observe a very interesting behaviour.

BDI: Patients who improved in their symptom severity did not improve in their BDI, while patients who worsened/remained improved in their BDI.

STAI-1: Patients who improved in their symptom severity did not improve in their STAI, while patients who worsened/remained worsened also in their STAI-1

What does this tell us about the CAR?

3.4 Temporal shift in CAR

Next, we stratify our CAR data based on clinical course after 8 months. We can see that improvement of the symptoms is associated with a temporal shift in the CAR.

We can look again at our delta scores, and we can see:

Patients who improved had

a higher value at wake-up at FUP, as depicted by a positive delta score

an earlier peak and stronger/faster flattening of the curve at FUP, as depicted by a negative delta score towards the end score

Patients who worsened/remained had

a lower value at wake-up at FUP, as depicted by a negative delta score

a later peak (i.e., temporal shift) at FUP, as depicted by a positive delta score towards the end

And this, ladies and gentlemen, is our “hidden” effect of symptom severity.

Why is this particularly interesting?

The CAR gets mostly triggered by the hippocampus and it may be linked to the hippocampus’ preparation of the HPA axis in anticipation of metabolic demands and stress.

## [1] "We play around with another threshold for symptom severity:"

## 
##      improved    same/worse same/worsened 
##            80            36             0

As a next step, we look at the statistics of the delta scores, and run post-hoc analyses on the individual timepoints to confirm that they significantly differ depending on clinical course. The post-hoc analyses might uncover the interaction, which cannot be found in the previous tests.

Indeed: An improvement in symptom severity is associated with with a temporal shift in the CAR.

## 
##   Simultaneous Tests for General Linear Hypotheses
## 
## Multiple Comparisons of Means: User-defined Contrasts
## 
## 
## Fit: lm(formula = sfmdrs ~ Cortisol * groups + bdi + contraception + 
##     age + psychMed, family = gaussian)
## 
## Linear Hypotheses:
##                               Estimate Std. Error t value Pr(>|t|)    
## (worse.1) - (improved.1) == 0    7.080      1.251   5.661 4.38e-08 ***
## (worse.2) - (improved.2) == 0    6.588      1.195   5.511 9.34e-08 ***
## (worse.3) - (improved.3) == 0    7.105      1.181   6.014 6.86e-09 ***
## (worse.4) - (improved.4) == 0    7.372      1.184   6.229 2.15e-09 ***
## (worse.5) - (improved.5) == 0    7.231      1.236   5.851 1.63e-08 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- none method)

3.5 Cortisol Amplitude

This analyses was suggested by Fabian. Basically, he said it could be interesting to look at the amplitude of cortisol throughout the day. So we basically calculate the amplitude (height) between the lowest and the highest peak of the cortisol within one day.

## [1] "Amplitude"

Statistical Analyses:

Can the cortisol amplitude predict clinical course (group improved vs. same or worse)? No, it cannot. However, it can predict symptom severity at fup.

##                  Df Sum Sq Mean Sq F value  Pr(>F)   
## Session           1    0.1    0.12   0.015 0.90193   
## group             1    9.9    9.93   1.292 0.25859   
## sfmdrs            1   65.7   65.65   8.545 0.00434 **
## DurationSymptoms  1    4.5    4.47   0.582 0.44733   
## gender            1   22.4   22.39   2.914 0.09113 . 
## smoke             1    2.7    2.72   0.354 0.55329   
## psychMed          1   13.4   13.42   1.747 0.18948   
## menopause         1    0.1    0.14   0.019 0.89198   
## contraception     1   10.7   10.68   1.390 0.24139   
## age               1    4.7    4.71   0.614 0.43543   
## Session:group     1    0.4    0.39   0.050 0.82309   
## Residuals        94  722.2    7.68                   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

We previously identified a reduced hippocampal volume in FND patients, unrelated to the CAR, and suggested this to be a trait vulnerability factor for the disorder. Let’s have a look at brain data.

4 Structural Alterations in FND

As a next step, we look into structural alterations between the two timepoints. We first look at the data in spm, where we implemented a regression analysis with symptom severity as dependent variable.
Based on a first look in spm, it seems that changes in symptom severity are associated with changes in the

basal ganglia (caudate + putamen)

Nucleus Accumbens

posterior part of the hippocampus

insula

Figure 4

This analyses do not survive when adding BDI as a covariate.
To make any further statements on the effect of symptom severity, we have to extract the data for external analyes in R.

We extracted ROI-wise (based on AAL3) averaged data and run again a linear model using BDI (and STAI) as a covariate of no-interest, also we stratify directly in groups based on clinical course. I only looked at those ROIs in which we found significant differences associated to SFMDRS (no BDI cov) in spm.

There are no significant effects or interactions between groups, timepoint and brain volume.

When removing BDI and STAI as covariate, a trend can be found in the Nucleus Caudate (does not survive post-hoc mcp):

## [1] "Results Regression Analysis of Brain Volume in R:"

##                      Df Sum Sq Mean Sq F value Pr(>F)  
## Session               1   0.06  0.0573   0.106 0.7450  
## sfmdrs                1   1.69  1.6882   3.134 0.0798 .
## group                 1   0.03  0.0287   0.053 0.8178  
## Session:sfmdrs        1   0.03  0.0264   0.049 0.8254  
## Session:group         1   0.08  0.0821   0.152 0.6970  
## sfmdrs:group          1   0.78  0.7817   1.451 0.2312  
## Session:sfmdrs:group  1   0.03  0.0302   0.056 0.8133  
## Residuals            98  52.78  0.5386                 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

## 
##   Simultaneous Tests for General Linear Hypotheses
## 
## Multiple Comparisons of Means: User-defined Contrasts
## 
## 
## Fit: glm(formula = Caud ~ groups, family = gaussian)
## 
## Linear Hypotheses:
##                                                                    Estimate
## (worse.fup) - (worse.t1) == 0                                       0.05817
## (improved.fup) - (improved.t1) == 0                                 0.02177
## (worse.fup) - (improved.fup) == 0                                  -0.04058
## (worse.t1) - (improved.t1) == 0                                    -0.07699
## ((worse.fup) - (worse.t1)) - ((improved.fup) - (improved.t1)) == 0  0.03640
##                                                                    Std. Error
## (worse.fup) - (worse.t1) == 0                                         0.17360
## (improved.fup) - (improved.t1) == 0                                   0.25263
## (worse.fup) - (improved.fup) == 0                                     0.21675
## (worse.t1) - (improved.t1) == 0                                       0.21675
## ((worse.fup) - (worse.t1)) - ((improved.fup) - (improved.t1)) == 0    0.30653
##                                                                    z value
## (worse.fup) - (worse.t1) == 0                                        0.335
## (improved.fup) - (improved.t1) == 0                                  0.086
## (worse.fup) - (improved.fup) == 0                                   -0.187
## (worse.t1) - (improved.t1) == 0                                     -0.355
## ((worse.fup) - (worse.t1)) - ((improved.fup) - (improved.t1)) == 0   0.119
##                                                                    Pr(>|z|)
## (worse.fup) - (worse.t1) == 0                                         0.738
## (improved.fup) - (improved.t1) == 0                                   0.931
## (worse.fup) - (improved.fup) == 0                                     0.851
## (worse.t1) - (improved.t1) == 0                                       0.722
## ((worse.fup) - (worse.t1)) - ((improved.fup) - (improved.t1)) == 0    0.905
## (Adjusted p values reported -- none method)

5 Genetic results

For our FND patients, we have genetic data on the following SNPs:

TPH2 rs4570625

TPH1 rs1800532

BDNF rs6265

BDNF rs1491850

DRD4 rs3758653

OXTR rs2254298

OXTR rs53576

DRD2 rs1799732

FKBP51 rs1360780

FKBP52 rs3800373

TPH1 rs1800532

We first do some quality controls:
1. Missing data: From one FND patients, no blood could be taken.
2. Hardy-Weinberg equilibrium: All data are in HWE

##           alleles major.allele.freq HWE      missing (%)
## rs1360780 C/T     74.5              0.710967 0          
## rs1491850 T/C     61.8              0.769792 0          
## rs1799732 G/I     87.3              1.000000 0          
## rs1800532 G/T     53.9              0.574052 0          
## rs2254298 G/A     91.2              1.000000 0          
## rs3758653 T/C     78.4              0.676882 0          
## rs3800373 A/C     74.5              0.710967 0          
## rs4570625 G/T     77.5              0.418175 0          
## rs53576   G/A     57.8              0.575946 0          
## rs6265    C/T     81.4              1.000000 0

Next, we will do an association study, based on our FND longitudinal data only. Previously, we compared FND patients with HC, and we could show that there is a significant association between:

rs53576 (OXTR) and the diagnosis of FND

In FND only:

rs4570625 (THP2) and symptom severity

rs1800532 (TPH1) and symptom severity

As a next step, we will look into genetic associations with:

clinical outcome

changes in brain volume

changes in cortisol (CAR)

5.1 Genetic association with clinical course in FND

We find a significant association between clinical course (group improved vs. worse/same) with the two SNPs for Brain-derived neurotrophic factor (BDNF):

rs1491850 (BDNF; dominant/co-dominant)

rs6265 (BDNF; dominant)

Age, gender, BDI, STAI were added as covariates of no-interest

##           comments codominant dominant recessive log-additive
## rs1360780        -    0.75428  0.53589   0.84770      0.68267
## rs1491850        -    0.02777  0.04313   0.51719      0.29831
## rs1799732        -    0.85524        -         -            -
## rs1800532        -    0.95337  0.75950   0.87838      0.78135
## rs2254298        -    0.14790        -         -            -
## rs3758653        -    0.24532  0.80424   0.12098      0.76650
## rs3800373        -    0.75428  0.53589   0.84770      0.68267
## rs4570625        -    0.78621  0.87900   0.51335      0.98062
## rs53576          -    0.69473  0.54000   0.44065      0.40900
## rs6265           -    0.05685  0.03834   0.71111      0.11889

We can also look at the individual statistics of those genes found to be associated with clinical course.

rs1491850 (BDNF):

In our results, risk allele carriers were associated with a poorer clinical outcome.

BDNF SNP associated with therapeutic response in OCD Ref and MDD Ref

BDNF Genotype associated with hippocampal volume after TBI Ref

rs6265 (BDNF):

In our results, risk allele carriers were associated with a poorer clinical outcome.

BDNF Polymorphism associated with error-specific networks Ref

lBDNF polymorphism associated with differrences in motor system function, altered short-term plasticiy and greater error in motor learning Ref

## [1] "Results rs1491850 T>C (BDNF):"

## 
## SNP: rs1491850  adjusted by: age gender bdi stai 
##            0    %  1    %   OR lower upper p-value  AIC
## Codominant                                             
## T/T        9 56.2 11 31.4 1.00             0.02777 64.2
## C/T        3 18.8 20 57.1 8.99  1.41 57.24             
## C/C        4 25.0  4 11.4 1.32  0.20  8.77             
## Dominant                                               
## T/T        9 56.2 11 31.4 1.00             0.04313 65.3
## C/T-C/C    7 43.8 24 68.6 4.21  0.99 17.98

## [1] "Results rs6265 G>A (BDNF):"

## 
## SNP: rs6265  adjusted by: age gender bdi stai 
##             0    %  1    %   OR lower upper p-value  AIC
## Codominant                                              
## C/C        13 81.2 21 60.0 1.00             0.05685 65.6
## C/T         2 12.5 13 37.1 8.93  1.05 75.90             
## T/T         1  6.2  1  2.9 0.95  0.04 23.66             
## Dominant                                                
## C/C        13 81.2 21 60.0 1.00             0.03834 65.1
## C/T-T/T     3 18.8 14 40.0 5.39  0.92 31.71

5.2 Genetic association with cortisol in FND

We find no significant association between the clinical course (group improved vs. worse/same) and changes in the CAR. SNP rs4570625 (TPH2) almost reaches significance. (Delta scores on AUC were calculated).

##           comments codominant dominant recessive log-additive
## rs1360780        -    0.97010  0.88250   0.90152      0.94943
## rs1491850        -    0.23653  0.20076   0.12079      0.09829
## rs1799732        -    0.55401        -         -            -
## rs1800532        -    0.68362  0.44190   0.89146      0.68423
## rs2254298        -    0.46565        -         -            -
## rs3758653        -    0.26084  0.11848   0.98506      0.20912
## rs3800373        -    0.97010  0.88250   0.90152      0.94943
## rs4570625        -    0.05671  0.21378   0.07855      0.48689
## rs53576          -    0.99521  0.93220   0.98555      0.96241
## rs6265           -    0.27983  0.10944   0.49812      0.11996

5.3 Genetic association with brain volume in FND

We first sum the brain volumes of some selected regions (for no specific reasons, but reduces the amount of comparisons to make, also in spm was very homogeneous across brain hemispheres). We take the following regions:

Insula

Hippocampus

Amygdala

Nucleus Caudate

Putamen

Nucleus Accumbens

For all analyses we take delta scores, which means that the statistics compares changes in brain volumes regarding clinical outcome.

## [1] "Genetic association with total insular volume:"

##           comments codominant dominant recessive log-additive
## rs1360780        -    0.05280  0.22629   0.01604      0.05422
## rs1491850        -    0.25108  0.09909   0.62363      0.17655
## rs1799732        -    0.24798        -         -            -
## rs1800532        -    0.22532  0.09945   0.27713      0.09377
## rs2254298        -    0.22827        -         -            -
## rs3758653        -    0.43789  0.20814   0.88565      0.28591
## rs3800373        -    0.05280  0.22629   0.01604      0.05422
## rs4570625        -    0.03168  0.87319   0.00955      0.61596
## rs53576          -    0.08742  0.05257   0.09389      0.02653
## rs6265           -    0.38340  0.68632   0.16406      0.41271

## [1] "Genetic association with total hippocampal volume:"

##           comments codominant dominant recessive log-additive
## rs1360780        -    0.87445  0.86793   0.60297      0.73216
## rs1491850        -    0.05754  0.01719   0.43926      0.04759
## rs1799732        -    0.39373        -         -            -
## rs1800532        -    0.97551  0.99465   0.83274      0.89856
## rs2254298        -    0.12637        -         -            -
## rs3758653        -    0.09066  0.83311   0.03263      0.33172
## rs3800373        -    0.87445  0.86793   0.60297      0.73216
## rs4570625        -    0.12602  0.12295   0.27001      0.25247
## rs53576          -    0.02466  0.03778   0.01818      0.00689
## rs6265           -    0.12398  0.33241   0.04292      0.13081

## [1] "Genetic association with total amygdalar volume:"

##           comments codominant dominant recessive log-additive
## rs1360780        -    0.36427  0.38617   0.16977      0.21661
## rs1491850        -    0.10031  0.09379   0.66953      0.36605
## rs1799732        -    0.74658        -         -            -
## rs1800532        -    0.28096  0.10949   0.61741      0.19542
## rs2254298        -    0.36827        -         -            -
## rs3758653        -    0.88829  0.94164   0.63071      0.80845
## rs3800373        -    0.36427  0.38617   0.16977      0.21661
## rs4570625        -    0.08306  0.52551   0.04155      0.94168
## rs53576          -    0.02416  0.01024   0.07806      0.00712
## rs6265           -    0.39222  0.56347   0.17126      0.33875

## [1] "Genetic association with total Ncl. caudate volume:"

##           comments codominant dominant recessive log-additive
## rs1360780        -    0.36488  0.76535   0.23753      0.80099
## rs1491850        -    0.05961  0.03983   0.89643      0.19344
## rs1799732        -    0.10315        -         -            -
## rs1800532        -    0.59712  0.99573   0.34218      0.58280
## rs2254298        -    0.82157        -         -            -
## rs3758653        -    0.46444  0.21351   0.64105      0.23767
## rs3800373        -    0.36488  0.76535   0.23753      0.80099
## rs4570625        -    0.00000  0.66270   0.00000      0.47375
## rs53576          -    0.05558  0.02230   0.12962      0.01831
## rs6265           -    0.53657  0.44880   0.59697      0.65830

## [1] "Genetic association with total putamen volume:"

##           comments codominant dominant recessive log-additive
## rs1360780        -    0.05018  0.01548   0.20845      0.01653
## rs1491850        -    0.76014  0.46445   0.83950      0.55314
## rs1799732        -    0.28991        -         -            -
## rs1800532        -    0.06151  0.01817   0.54057      0.06732
## rs2254298        -    0.91599        -         -            -
## rs3758653        -    0.50489  0.38359   0.65451      0.59658
## rs3800373        -    0.05018  0.01548   0.20845      0.01653
## rs4570625        -    0.07406  0.16383   0.05085      0.06942
## rs53576          -    0.81423  0.71522   0.71727      0.96578
## rs6265           -    0.84410  0.55879   0.82601      0.57417

## [1] "Genetic association with total Ncl. accumbens volume:"

##           comments codominant dominant recessive log-additive
## rs1360780        -    0.68373  0.66444   0.58570      0.91190
## rs1491850        -    0.83659  0.82082   0.69258      0.96517
## rs1799732        -    0.17142        -         -            -
## rs1800532        -    0.14089  0.69329   0.04969      0.16638
## rs2254298        -    0.76759        -         -            -
## rs3758653        -    0.84055  0.56459   0.95198      0.62626
## rs3800373        -    0.68373  0.66444   0.58570      0.91190
## rs4570625        -    0.00011  0.50079   0.00002      0.09588
## rs53576          -    0.02990  0.00792   0.45857      0.03167
## rs6265           -    0.96232  0.80961   0.95724      0.85629

5.3.1 Insula x Genetics

The change of the volume in the insula is associated with:

rs1360780 (FKBP5): Only few major allele carrier. I would not consider this result as being reliable.

rs3800373 (FKBP5): Again, only few major allele carrier.

rs4570625 (TPH2): There is only one homozygote for minor allele. For codominant model (heterozygote), risk allele carriers had smaller changes in brain volume in the Insula.

## [1] "Results rs1360780 C>T (FKBP5; recessive model):"

## 
## SNP: rs1360780  adjusted by: group bdi stai 
##            n       me      se     dif  lower   upper p-value  AIC
## Recessive                                                        
## C/C-C/T   47  0.01042 0.06076  0.0000                0.01604 68.6
## T/T        4 -0.58105 0.35608 -0.5935 -1.059 -0.1282

## [1] "Results rs3800373 T>G (FKBP5; recessive model):"

## 
## SNP: rs3800373  adjusted by: group bdi stai 
##            n       me      se     dif  lower   upper p-value  AIC
## Recessive                                                        
## A/A-C/A   47  0.01042 0.06076  0.0000                0.01604 68.6
## C/C        4 -0.58105 0.35608 -0.5935 -1.059 -0.1282

## [1] "Results rs4570625 G>T (TPH2; recessive/codominant model):"

## 
## SNP: rs4570625  adjusted by: group bdi stai 
##             n       me      se      dif   lower  upper  p-value  AIC
## Codominant                                                          
## G/G        29 -0.03219 0.09638  0.00000                0.031681 69.3
## G/T        21 -0.09230 0.06956 -0.06717 -0.3238 0.1894              
## T/T         1  1.03710 0.00000  1.21382  0.2894 2.1382              
## Recessive                                                           
## G/G-G/T    50 -0.05743 0.06266  0.00000                0.009551 67.6
## T/T         1  1.03710 0.00000  1.25065  0.3444 2.1569

5.3.2 Hippocampus x Genetics

The change of the volume in the hippocampus is associated with:

rs1491850 (BDNF): Risk allele carriers had a bigger change in volume in the hippocampus.

rs3758653 (DRD4): There is only three homozygote for minor allele. Results should not be considered reliable.

rs53576 (OXTR): Risk allele carriers had a smaller change in volume in the hippocampus.

## [1] "Results rs1491850 G>T (BDNF; dominant model):"

## 
## SNP: rs1491850  adjusted by: group bdi stai 
##           n       me      se     dif   lower  upper p-value   AIC
## Dominant                                                         
## T/T      20 -0.02800 0.03452 0.00000                0.01719 -53.1
## C/T-C/C  31  0.08645 0.02172 0.09899 0.02051 0.1775

## [1] "Results rs3758653 T>C (DRD4; dominant model):"

## 
## SNP: rs3758653  adjusted by: group bdi stai 
##            n       me      se     dif   lower    upper p-value   AIC
## Recessive                                                           
## T/T-T/C   48  0.04967 0.02077  0.0000                  0.03263 -51.8
## C/C        3 -0.08800 0.06391 -0.1859 -0.3512 -0.02052

## [1] "Results rs53576 G>T (OXTR; (co)dominant model):"

## 
## SNP: rs53576  adjusted by: group bdi stai 
##             n       me      se      dif   lower     upper p-value   AIC
## Codominant                                                             
## G/G        18  0.08883 0.02250  0.00000                   0.02466 -53.1
## G/A        23  0.04016 0.03632 -0.05876 -0.1417  0.024167              
## A/A        10 -0.04026 0.03673 -0.15084 -0.2552 -0.046515              
## Dominant                                                               
## G/G        18  0.08883 0.02250  0.00000                   0.03778 -51.6
## G/A-A/A    33  0.01579 0.02810 -0.08639 -0.1656 -0.007228              
## Recessive                                                              
## G/G-G/A    41  0.06153 0.02272  0.00000                   0.01818 -53.0
## A/A        10 -0.04026 0.03673 -0.11778 -0.2120 -0.023531

5.3.3 Amygdala x Genetics

The change of the volume in the amygdala is associated with:

rs4570625 (TPH2): There is only one homozygote for minor allele. Results should not be considered reliable.

rs53576 (OXTR): Minor allele carrier had a smaller difference in amygdalar volume (no difference = no change)

## [1] "Results rs4570625 G>T (TPH2; recessive model):"

## 
## SNP: rs4570625  adjusted by: group bdi stai 
##            n      me      se    dif  lower  upper p-value   AIC
## Recessive                                                      
## G/G-G/T   50 0.02786 0.02265 0.0000               0.04155 -38.2
## T/T        1 0.37120 0.00000 0.3435 0.0224 0.6646

## [1] "Results rs53576 G>T (OXTR; (co)dominant model):"

## 
## SNP: rs53576  adjusted by: group bdi stai 
##             n        me      se     dif   lower     upper p-value AIC
## Codominant                                                           
## G/G        18  0.102283 0.04534  0.0000                   0.02416 -40
## G/A        23  0.010478 0.03253 -0.1031 -0.1974 -0.008842            
## A/A        10 -0.031780 0.02758 -0.1593 -0.2780 -0.040708            
## Dominant                                                             
## G/G        18  0.102283 0.04534  0.0000                   0.01024 -41
## G/A-A/A    33 -0.002327 0.02416 -0.1200 -0.2078 -0.032170

5.3.4 Ncl. Caudate x Genetics

The change of the volume in the Ncl. Caudate is associated with:

rs1491850 (BDNF): Risk allele carriers had a higher change in volume in the Ncl. Caudate

rs4570625 (THP2): Heterozygote risk allele carriers had a smaller change in volume in the Ncl. Caudate

rs53576 (OXTR): Risk allele carriers had a smaller change in volume in the Ncl. Caudate

## [1] "Results rs1491850 G>T (BDNF; dominant model):"

## 
## SNP: rs1491850  adjusted by: group bdi stai 
##           n       me      se     dif    lower  upper p-value   AIC
## Dominant                                                          
## T/T      20 -0.01050 0.02881 0.00000                 0.03983 -46.8
## C/T-C/C  31  0.08043 0.02679 0.09006 0.006624 0.1735

## [1] "Results rs4570625 G>T (THP2; codominant and recessive model):"

## 
## SNP: rs4570625  adjusted by: group bdi stai 
##             n      me      se      dif   lower   upper   p-value   AIC
## Codominant                                                            
## G/G        29 0.04758 0.02021  0.00000                 3.248e-06 -68.7
## G/T        21 0.00859 0.02505 -0.04392 -0.1102 0.02237                
## T/T         1 0.72320 0.00000  0.65959  0.4208 0.89842                
## Recessive                                                             
## G/G-G/T    50 0.03121 0.01583  0.00000                 1.017e-06 -68.9
## T/T         1 0.72320 0.00000  0.68368  0.4459 0.92148

## [1] "Results rs53576 G>T (OXTR; dominant model):"

## 
## SNP: rs53576  adjusted by: group bdi stai 
##           n      me      se      dif  lower    upper p-value AIC
## Dominant                                                        
## G/G      18 0.10399 0.04194  0.00000                  0.0223 -48
## G/A-A/A  33 0.01247 0.02059 -0.09896 -0.181 -0.01695

5.3.5 Putamen x Genetics

The change of the volume in the putamen is associated with:

rs1360780 (FKBP5): Risk allele carriers had a smaller change in volume in the Putamen

rs1800532 (THP1): Risk allele carriers had a bigger change in volume in the Putamen

rs3800373 (FKBP5): Risk allele carriers had a smaller change in volume in the Putamen

## [1] "Results rs1360780 C>T (FKBP5; dominant model):"

## 
## SNP: rs1360780  adjusted by: group bdi stai 
##           n       me     se     dif   lower    upper p-value   AIC
## Dominant                                                          
## C/C      29  0.08006 0.1083  0.0000                  0.01548 100.4
## C/T-T/T  22 -0.30904 0.1430 -0.4346 -0.7733 -0.09585

## [1] "Results rs1800532 C>A (TPH1; dominant model):"

## 
## SNP: rs1800532  adjusted by: group bdi stai 
##           n       me      se    dif   lower  upper p-value   AIC
## Dominant                                                        
## G/G      16 -0.33944 0.21052 0.0000                0.01817 100.8
## G/T-T/T  35  0.02726 0.08576 0.4629 0.09253 0.8332

## [1] "Results rs3800373 T>G (FKBP5; dominant model):"

## 
## SNP: rs3800373  adjusted by: group bdi stai 
##           n       me     se     dif   lower    upper p-value   AIC
## Dominant                                                          
## A/A      29  0.08006 0.1083  0.0000                  0.01548 100.4
## C/A-C/C  22 -0.30904 0.1430 -0.4346 -0.7733 -0.09585

5.3.6 Ncl. Accumbens x Genetics

The change of the volume in the Ncl. Accumbens is associated with:

rs1800532 (THP1): Risk allele carriers had a bigger change in volume in the Ncl. Accumbens

rs4570625 (TPH2): Heterozygote risk allele carriers had a bigger change in volume in the Ncl. Accumbens

rs3800373 (FKBP5): Risk allele carriers had a smaller change in volume in the Ncl. Accumbens

## [1] "Results rs1800532 C>A (TPH1; recessive model):"

## 
## SNP: rs1800532  adjusted by: group bdi stai 
##            n       me       se    dif    lower  upper p-value   AIC
## Recessive                                                          
## G/G-G/T   39 0.003256 0.009456 0.0000                 0.04969 -97.5
## T/T       12 0.058825 0.041212 0.0587 0.001625 0.1158

## [1] "Results rs4570625 G>T (DRD4; codominant and recessive model):"

## 
## SNP: rs4570625  adjusted by: group bdi stai 
##             n       me       se      dif    lower   upper   p-value    AIC
## Codominant                                                                
## G/G        29 0.006921 0.008831 0.000000                  1.056e-04 -111.9
## G/T        21 0.012667 0.021627 0.004932 -0.03847 0.04833                 
## T/T         1 0.366200 0.000000 0.377125  0.22077 0.53349                 
## Recessive                                                                 
## G/G-G/T    50 0.009334 0.010305 0.000000                  1.723e-05 -113.9
## T/T         1 0.366200 0.000000 0.374421  0.22149 0.52735

## [1] "Results rs53576 G>T (OXTR; (co)dominant model):"

## 
## SNP: rs53576  adjusted by: group bdi stai 
##             n        me       se      dif   lower    upper  p-value    AIC
## Codominant                                                                
## G/G        18  0.059311 0.029491  0.00000                  0.029903  -99.2
## G/A        23 -0.010200 0.009441 -0.07110 -0.1239 -0.01830                
## A/A        10 -0.000010 0.015522 -0.06409 -0.1305  0.00234                
## Dominant                                                                  
## G/G        18  0.059311 0.029491  0.00000                  0.007917 -101.1
## G/A-A/A    33 -0.007112 0.007996 -0.06900 -0.1177 -0.02030

5.3.7 Summary Genetics x Brain Volume

In summary, a consistent pattern of association with brain changes was found. As such, risk allele carriers of:

BDNF and THP1 were associated with increased changes in volumes

FKBP5, TPH2 and OXTR were associated with decreased changes in volumes

6 Prognostic Markers Models

For the last chapter, we will work on some predictive models, where we will use data from the fup as outcome variables, and try to predict them based on the variables at t1. As genotype did not change between sessions (or at least we hope our participants did not undergo gene-modifying therapies), we used sfmdrs at t1 as a covariate as well.
There are not many findings, though. Here some of the most important findings:

1. Cortisol amplitude at T1 can predict symptom severity at fup
2. CARi/PACC/DBCC at T1 does not predict symptom severity at fup
3. Brain volume at T1 does not predict symptom severity at fup
4. TPH1, and OXTR polymorphism was associated with symptom severity at fup (This analyses are not really important, as genotype did not change between sessions). Also not always a reliable number of risk allele carriers. Most important findings are:

TPH1 (rs1800532): risk allele carriers associated with increased symptom severity at fup.

THP2 (rs4570625): Heterozygous risk allele carriers associated with reduced symptom severity at fup (only one homozygote risk allele carrier though).

## [1] "Cortisol Amplitude at t1, predicts symptom severity at fup:"

## 
## Call:
## glm(formula = sfmdrs[Data.AD$Session == "fup"] ~ Amplitude[Data.AD$Session == 
##     "t1"] + age[Data.AD$Session == "t1"] + gender[Data.AD$Session == 
##     "t1"] + DurationSymptoms[Data.AD$Session == "t1"] + bdi[Data.AD$Session == 
##     "t1"] + stai1[Data.AD$Session == "t1"], data = Data.AD)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -16.163   -6.739   -1.053    4.655   29.892  
## 
## Coefficients:
##                                           Estimate Std. Error t value Pr(>|t|)
## (Intercept)                                6.14274    7.07474   0.868   0.3898
## Amplitude[Data.AD$Session == "t1"]         1.24785    0.57032   2.188   0.0338
## age[Data.AD$Session == "t1"]               0.06460    0.11548   0.559   0.5786
## gender[Data.AD$Session == "t1"]male        3.47514    3.30153   1.053   0.2980
## DurationSymptoms[Data.AD$Session == "t1"] -0.03072    0.02763  -1.112   0.2719
## bdi[Data.AD$Session == "t1"]               0.29872    0.17438   1.713   0.0934
## stai1[Data.AD$Session == "t1"]            -0.27388    0.16257  -1.685   0.0988
##                                            
## (Intercept)                                
## Amplitude[Data.AD$Session == "t1"]        *
## age[Data.AD$Session == "t1"]               
## gender[Data.AD$Session == "t1"]male        
## DurationSymptoms[Data.AD$Session == "t1"]  
## bdi[Data.AD$Session == "t1"]              .
## stai1[Data.AD$Session == "t1"]            .
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for gaussian family taken to be 103.9779)
## 
##     Null deviance: 6036.5  on 52  degrees of freedom
## Residual deviance: 4783.0  on 46  degrees of freedom
## AIC: 405.04
## 
## Number of Fisher Scoring iterations: 2

## 
## Call:
## glm(formula = sfmdrs[Data.AD$Session == "fup"] ~ Amplitude[Data.AD$Session == 
##     "t1"] * group[Data.AD$Session == "fup"] + age[Data.AD$Session == 
##     "t1"] + gender[Data.AD$Session == "t1"] + DurationSymptoms[Data.AD$Session == 
##     "t1"] + bdi[Data.AD$Session == "t1"] + stai1[Data.AD$Session == 
##     "t1"], data = Data.AD)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -21.821   -5.441   -1.147    4.739   26.548  
## 
## Coefficients:
##                                                                              Estimate
## (Intercept)                                                                  15.95216
## Amplitude[Data.AD$Session == "t1"]                                           -1.19067
## group[Data.AD$Session == "fup"]same/worse                                    -7.91059
## age[Data.AD$Session == "t1"]                                                  0.12209
## gender[Data.AD$Session == "t1"]male                                           2.15932
## DurationSymptoms[Data.AD$Session == "t1"]                                    -0.01952
## bdi[Data.AD$Session == "t1"]                                                  0.39161
## stai1[Data.AD$Session == "t1"]                                               -0.42241
## Amplitude[Data.AD$Session == "t1"]:group[Data.AD$Session == "fup"]same/worse  2.74023
##                                                                              Std. Error
## (Intercept)                                                                    10.81439
## Amplitude[Data.AD$Session == "t1"]                                              1.41639
## group[Data.AD$Session == "fup"]same/worse                                       8.44541
## age[Data.AD$Session == "t1"]                                                    0.11214
## gender[Data.AD$Session == "t1"]male                                             3.17989
## DurationSymptoms[Data.AD$Session == "t1"]                                       0.02658
## bdi[Data.AD$Session == "t1"]                                                    0.16938
## stai1[Data.AD$Session == "t1"]                                                  0.16751
## Amplitude[Data.AD$Session == "t1"]:group[Data.AD$Session == "fup"]same/worse    1.55262
##                                                                              t value
## (Intercept)                                                                    1.475
## Amplitude[Data.AD$Session == "t1"]                                            -0.841
## group[Data.AD$Session == "fup"]same/worse                                     -0.937
## age[Data.AD$Session == "t1"]                                                   1.089
## gender[Data.AD$Session == "t1"]male                                            0.679
## DurationSymptoms[Data.AD$Session == "t1"]                                     -0.734
## bdi[Data.AD$Session == "t1"]                                                   2.312
## stai1[Data.AD$Session == "t1"]                                                -2.522
## Amplitude[Data.AD$Session == "t1"]:group[Data.AD$Session == "fup"]same/worse   1.765
##                                                                              Pr(>|t|)
## (Intercept)                                                                    0.1473
## Amplitude[Data.AD$Session == "t1"]                                             0.4051
## group[Data.AD$Session == "fup"]same/worse                                      0.3540
## age[Data.AD$Session == "t1"]                                                   0.2822
## gender[Data.AD$Session == "t1"]male                                            0.5007
## DurationSymptoms[Data.AD$Session == "t1"]                                      0.4666
## bdi[Data.AD$Session == "t1"]                                                   0.0255
## stai1[Data.AD$Session == "t1"]                                                 0.0154
## Amplitude[Data.AD$Session == "t1"]:group[Data.AD$Session == "fup"]same/worse   0.0845
##                                                                               
## (Intercept)                                                                   
## Amplitude[Data.AD$Session == "t1"]                                            
## group[Data.AD$Session == "fup"]same/worse                                     
## age[Data.AD$Session == "t1"]                                                  
## gender[Data.AD$Session == "t1"]male                                           
## DurationSymptoms[Data.AD$Session == "t1"]                                     
## bdi[Data.AD$Session == "t1"]                                                 *
## stai1[Data.AD$Session == "t1"]                                               *
## Amplitude[Data.AD$Session == "t1"]:group[Data.AD$Session == "fup"]same/worse .
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for gaussian family taken to be 93.78134)
## 
##     Null deviance: 6036.5  on 52  degrees of freedom
## Residual deviance: 4126.4  on 44  degrees of freedom
## AIC: 401.22
## 
## Number of Fisher Scoring iterations: 2

## [1] "Genetic association with symptom severity at fup:"

##           comments codominant dominant recessive log-additive
## rs1360780        -    0.74460  0.82520   0.54101      0.95302
## rs1491850        -    0.45615  0.71291   0.36548      0.84222
## rs1799732        -    0.12766        -         -            -
## rs1800532        -    0.03957  0.39331   0.01058      0.04139
## rs2254298        -    0.55284        -         -            -
## rs3758653        -    0.59619  0.83795   0.38659      0.87464
## rs3800373        -    0.74460  0.82520   0.54101      0.95302
## rs4570625        -    0.00038  0.97192   0.00007      0.30853
## rs53576          -    0.18958  0.07489   0.30863      0.08062
## rs6265           -    0.80279  0.54259   0.64988      0.50627

## [1] "Results rs1800532 C>A (TPH1; dominant, recessive model):"

## 
## SNP: rs1800532  adjusted by: group bdi stai sfmdrs 
##            n     me    se   dif  lower  upper p-value   AIC
## Dominant                                                   
## G/G       16  9.188 3.273 0.000               0.39331 380.9
## G/T-T/T   35  8.800 1.706 2.580 -3.287  8.446              
## Recessive                                                  
## G/G-G/T   39  7.333 1.587 0.000               0.01058 374.3
## T/T       12 14.083 3.781 7.908  2.098 13.718

## [1] "Results rs4570625 G>T (THP2; codominant and recessive model):"

## 
## SNP: rs4570625  adjusted by: group bdi stai 
##             n     me    se    dif  lower  upper  p-value   AIC
## Codominant                                                    
## G/G        29  9.655 1.762  0.000               0.002728 383.2
## G/T        21  6.143 2.162 -4.064 -9.635  1.508               
## T/T         1 46.000 0.000 32.034 11.963 52.105               
## Recessive                                                     
## G/G-G/T    50  8.180 1.375  0.000               0.001634 383.5
## T/T         1 46.000 0.000 34.262 14.201 54.323

Session info

## R version 4.2.1 (2022-06-23)
## Platform: x86_64-apple-darwin17.0 (64-bit)
## Running under: macOS Big Sur ... 10.16
## 
## Matrix products: default
## BLAS:   /Library/Frameworks/R.framework/Versions/4.2/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/4.2/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] exactRankTests_0.8-35 SNPassoc_2.0-17       Gmisc_3.0.1          
##  [4] htmlTable_2.4.1       Rcpp_1.0.9            gridExtra_2.3        
##  [7] multcomp_1.4-20       TH.data_1.1-1         MASS_7.3-58.1        
## [10] survival_3.4-0        mvtnorm_1.1-3         lawstat_3.5          
## [13] car_3.1-1             carData_3.0-5         writexl_1.4.2        
## [16] maditr_0.8.3          reshape2_1.4.4        rstatix_0.7.1        
## [19] plotly_4.10.1         plyr_1.8.8            cowplot_1.1.1        
## [22] gridGraphics_0.5-1    gg.gap_1.3            ggpubr_0.5.0         
## [25] wesanderson_0.3.6     RColorBrewer_1.1-3    xlsx_0.6.5           
## [28] readxl_1.4.1          forcats_0.5.2         stringr_1.5.0        
## [31] dplyr_1.0.10          purrr_0.3.5           readr_2.1.3          
## [34] tidyr_1.2.1           tibble_3.1.8          ggplot2_3.4.0        
## [37] tidyverse_1.3.2       MuMIn_1.47.1          lmerTest_3.1-3       
## [40] lme4_1.1-30           Matrix_1.5-1         
## 
## loaded via a namespace (and not attached):
##   [1] utf8_1.2.2                             
##   [2] rms_6.3-0                              
##   [3] tidyselect_1.2.0                       
##   [4] AnnotationDbi_1.58.0                   
##   [5] RSQLite_2.2.18                         
##   [6] htmlwidgets_1.5.4                      
##   [7] BiocParallel_1.30.4                    
##   [8] munsell_0.5.0                          
##   [9] codetools_0.2-18                       
##  [10] interp_1.1-3                           
##  [11] withr_2.5.0                            
##  [12] colorspace_2.0-3                       
##  [13] filelock_1.0.2                         
##  [14] Biobase_2.56.0                         
##  [15] highr_0.9                              
##  [16] knitr_1.40                             
##  [17] rstudioapi_0.14                        
##  [18] stats4_4.2.1                           
##  [19] ggsignif_0.6.4                         
##  [20] rJava_1.0-6                            
##  [21] MatrixGenerics_1.8.1                   
##  [22] Rdpack_2.4                             
##  [23] labeling_0.4.2                         
##  [24] GenomeInfoDbData_1.2.8                 
##  [25] bit64_4.0.5                            
##  [26] farver_2.1.1                           
##  [27] vctrs_0.5.0                            
##  [28] generics_0.1.3                         
##  [29] xfun_0.37                              
##  [30] BiocFileCache_2.4.0                    
##  [31] timechange_0.1.1                       
##  [32] R6_2.5.1                               
##  [33] GenomeInfoDb_1.32.4                    
##  [34] Kendall_2.2.1                          
##  [35] DelayedArray_0.22.0                    
##  [36] bitops_1.0-7                           
##  [37] cachem_1.0.6                           
##  [38] assertthat_0.2.1                       
##  [39] BiocIO_1.6.0                           
##  [40] scales_1.2.1                           
##  [41] nnet_7.3-18                            
##  [42] googlesheets4_1.0.1                    
##  [43] gtable_0.3.1                           
##  [44] sandwich_3.0-2                         
##  [45] rlang_1.0.6                            
##  [46] MatrixModels_0.5-1                     
##  [47] splines_4.2.1                          
##  [48] rtracklayer_1.56.1                     
##  [49] lazyeval_0.2.2                         
##  [50] gargle_1.2.1                           
##  [51] broom_1.0.1                            
##  [52] checkmate_2.1.0                        
##  [53] BiocManager_1.30.18                    
##  [54] yaml_2.3.6                             
##  [55] abind_1.4-5                            
##  [56] modelr_0.1.9                           
##  [57] GenomicFeatures_1.48.4                 
##  [58] backports_1.4.1                        
##  [59] Hmisc_4.7-1                            
##  [60] tools_4.2.1                            
##  [61] ellipsis_0.3.2                         
##  [62] jquerylib_0.1.4                        
##  [63] BiocGenerics_0.42.0                    
##  [64] progress_1.2.2                         
##  [65] base64enc_0.1-3                        
##  [66] zlibbioc_1.42.0                        
##  [67] poisbinom_1.0.1                        
##  [68] RCurl_1.98-1.9                         
##  [69] prettyunits_1.1.1                      
##  [70] rpart_4.1.19                           
##  [71] deldir_1.0-6                           
##  [72] S4Vectors_0.34.0                       
##  [73] zoo_1.8-11                             
##  [74] SummarizedExperiment_1.26.1            
##  [75] haven_2.5.1                            
##  [76] cluster_2.1.4                          
##  [77] fs_1.5.2                               
##  [78] magrittr_2.0.3                         
##  [79] data.table_1.14.6                      
##  [80] forestplot_3.1.1                       
##  [81] SparseM_1.81                           
##  [82] reprex_2.0.2                           
##  [83] googledrive_2.0.0                      
##  [84] matrixStats_0.62.0                     
##  [85] hms_1.1.2                              
##  [86] xlsxjars_0.6.1                         
##  [87] evaluate_0.17                          
##  [88] XML_3.99-0.11                          
##  [89] jpeg_0.1-9                             
##  [90] IRanges_2.30.1                         
##  [91] biomaRt_2.52.0                         
##  [92] compiler_4.2.1                         
##  [93] crayon_1.5.2                           
##  [94] arsenal_3.6.3                          
##  [95] minqa_1.2.5                            
##  [96] htmltools_0.5.3                        
##  [97] mgcv_1.8-41                            
##  [98] tzdb_0.3.0                             
##  [99] Formula_1.2-4                          
## [100] lubridate_1.9.0                        
## [101] DBI_1.1.3                              
## [102] dbplyr_2.2.1                           
## [103] rappdirs_0.3.3                         
## [104] boot_1.3-28                            
## [105] BiocStyle_2.24.0                       
## [106] cli_3.4.1                              
## [107] rbibutils_2.2.13                       
## [108] parallel_4.2.1                         
## [109] TxDb.Hsapiens.UCSC.hg19.knownGene_3.2.2
## [110] GenomicRanges_1.48.0                   
## [111] pkgconfig_2.0.3                        
## [112] GenomicAlignments_1.32.1               
## [113] haplo.stats_1.8.9                      
## [114] numDeriv_2016.8-1.1                    
## [115] foreign_0.8-83                         
## [116] xml2_1.3.3                             
## [117] bslib_0.4.0                            
## [118] XVector_0.36.0                         
## [119] rvest_1.0.3                            
## [120] VariantAnnotation_1.42.1               
## [121] digest_0.6.30                          
## [122] Biostrings_2.64.1                      
## [123] rmarkdown_2.20                         
## [124] cellranger_1.1.0                       
## [125] restfulr_0.0.15                        
## [126] curl_4.3.3                             
## [127] Rsamtools_2.12.0                       
## [128] quantreg_5.94                          
## [129] rjson_0.2.21                           
## [130] nloptr_2.0.3                           
## [131] lifecycle_1.0.3                        
## [132] nlme_3.1-160                           
## [133] jsonlite_1.8.3                         
## [134] BSgenome_1.64.0                        
## [135] viridisLite_0.4.1                      
## [136] fansi_1.0.3                            
## [137] pillar_1.8.1                           
## [138] lattice_0.20-45                        
## [139] KEGGREST_1.36.3                        
## [140] fastmap_1.1.0                          
## [141] httr_1.4.4                             
## [142] glue_1.6.2                             
## [143] png_0.1-7                              
## [144] bit_4.0.4                              
## [145] stringi_1.7.8                          
## [146] sass_0.4.2                             
## [147] blob_1.2.3                             
## [148] polspline_1.1.20                       
## [149] org.Hs.eg.db_3.15.0                    
## [150] memoise_2.0.1                          
## [151] latticeExtra_0.6-30

FND goes Longitudinal: Cortisol, Brain Structure, and genetic predisposition in FND

Samantha Weber

20/03/23